Research On Low Signal-To-Noise Ratio Signal Detection Based On Non-Linear Theory

Detecting unknown signals is an important part of the communication countermeasure. How to get the characters of the unknown signals in a finite time is becoming a hot topic. The transmission signal is usually with high frequency, which is at the range of 3MHz to 3000MHz. Furthermore, people transmit the useful signals with low Signal-to-Noise Ratio(SNR) to avoid the detection from the enemy. The traditional linear signal detection methods are usually time-based and can not detect the signal in a finite time. SNR lower than -10dB can not be detected with those methods. Thus, new methods should be used in the condition.In the scenario above, the research focuses on how to detect the high frequency low SNR signals with non-linear methods. Duffing chaotic oscillator is a well-known non-linear model. It is sensitive to the weak signals but immune to noise. This unique property make it possible to detect the low SNR signals in communication area. The wavelet transform method is good at signal de-noising and can increase the SNR. Based on the two methods, a combined non-linear model is used to detect the high frequency low SNR signals.The main research and contributions of this paper are as follows:(1) The mathematical structure and system properties of the non-linear model are analysed. The criterion of the system state is selected, and detection threshold is calculated based on the experiment datas.(2) The detection ability of the non-linear model on high frequency low SNR sine signals is researched. The frequency of the unknown sine signals can be detected. The detectable frequency is in an order of 100MHz,and the lowest detectable SNR is about -30dB. The amplitude of the unknown sine signals can be detected in an order of 10-3, and the measurement accuracy decreases as the amplitude reduces.(3) The detection possbility of the non-linear model on BPSK signals is discussed. A non-linear model group is used to detect the BPSK signals. The results show that a BPSK signal can be detected with a frequency range from 225MHz to 400MHz, and the detectable SNR is about -20 dB.